liblsl 0.7.0

A Dart (and Flutter) native library for working with Lab Streaming Layer (LSL / liblsl).

0.7.0

This release is a major update that includes breaking changes. This update provides a large performance improvement by reusing a buffer for samples, reducing the number of allocations and copies required when sending samples.

There are also new packages that complement liblsl.dart, and are still work-in-progress, but will allow setting up an entire experiment workflow without requiring any programming. Currently these packages include:

• liblsl_timing: This package provides an application for measuring LSL timing performance with your specific device and network configuration. The test app automatically coordinates between all connected devices on the network, and passes the test configuration via a coordinator - the coordinator is the first device that starts the application. There are three tests included:
  • Latency: This test measures the latency of sending and receiving samples over LSL at the specified frequency.
  • Sync: This test is intended to measure the clock synchronization and drift between devices.
  • Interactive: This test has a button on screen and when the button is pressed, an LSL sample is sent, all receiving devices will then flash a black square on the screen. This test is intended to measure the entire end-to-end latency, including Flutter rendering, input and display lag. To measure this effectively, it would be ideal to have a touch sensor of some kind (e.g. FSR) and a photodiode sensor to measure the moment of touch and the moment of display change, respectively. There is a companion app in C++ written for a Bela (Beaglebone Black) which takes digital inputs and logs the timestamps along with the LSL samples. You can see more at the bela-lsl-timing repository.
• liblsl_analysis: This package allows for analysis of the data provided by the liblsl_timing package. It currently only allows loading a single TSV file from the timing test, but will be updated to allow loading files from all the reporting devices to create a comprehensive report of the timing performance across all devices. The report will include:
  • Latency statistics for each device
  • Clock synchronization and drift statistics
  • Interactive test results with timestamps and latency measurements (if available)
• Use custom fork of liblsl which allows API configuration to be specified at runtime (once, before any other LSL functions are called). This means that anything in the LSL API configuration file can be set, including on mobile platforms that do not support environment variables or allow editing files in /etc or ./.
  • This is exposed via the C/C++ API as lsl_set_config_filename and lsl_set_config_content to set the configuration file name and content (directly as a std::string/char*), respectively. The Dart API now provides a LSLConfig class that can be used to set the configuration file name and content, which can be used in LSL.setConfigFilename and LSL.setConfigContent methods.
• New LSLReusableBuffer class for allowing sample structures to avoid creating new instances for each sample. This reuse significantly enhances the performance by reducing the allocations during sample pulling and pushing.
• Generic push sample functions have now been replaced with specific implementations for each type e.g. LslPushSampleFloat.
• A new helper function runPreciseInterval has been added to handle precision interval timing, using an adjustable busy-wait loop. The API is subject to change, but allows for a callback and a mutable dyanmic state to be passed in, which can be used to update the state of the callback. This is useful for implementing precise timing in LSL applications (such as requiring 1000Hz sample creation with high precision -> resulting in a mean of 1.0000, median of 1.0082 ms over 180,000 samples).
• hooks package updated from 0.19.0 to 0.19.1.
• native_toolchain_c updated from ^0.16.0 to ^0.16.1.
• ffigen updated from 18.1.0 to 19.0.0.
• test package updated from 1.25.15 to 1.26.0.

0.6.2-dev.0

• Make SDK constraint ^3.9.0-0

0.6.1

• Attempt to fix package issue on pub.dev
• bump dart SDK constraint to ^3.9.0 due to hooks package
• Added LSLStreamResolver mixin (empty for now).
• Added dartdoc dev dependency, and added doc directory to .gitignore

0.6.0

• New LSLIsolatedInlet.getTimeCorrection method to get the LSL reported sample time correction
• Removed deprecated LSLStreamInlet and LSLStreamOutlet classes
• Remove deprecated native_assets_cli dependency
• Add hooks 0.19.0 dependency instead of native_assets_cli
• Add code_assets 0.19.0
• Update native_toolchain_c to ^0.16.0

0.5.1

• Fix package name on Android
• Update ffigen to 18.1.0
• Update native_assets_cli to 0.14.0
• Update native_toolchain_c to 0.11.0

0.5.0

• Generated dylib is now without the lib prefix (if it is the prefix on the platform). i.e. libliblsl.so is now just liblsl.so
• Tested working on Raspberry Pi 4 (64bit)
• Tests are now less chatty

0.4.1

• Update native_assets_cli to 0.13.0
• Update native_toolchain_c to 0.10.0

0.4.0

• Inlet and outlets now pass sample pointer addresses rather than sample objects, making them more efficient (note: this has no public facing changes to the API)
• Updated the liblsl_test package (android NDK version, entitlements, manifests)
• Updated readme documentation, describing Android and iOS specifics

0.3.0

• Outlets and inlets now have to be destroyed by the user
• Outlets and inlets are now contained in Isolates
• Updated liblsl to 7e61a2e
• Added a fully self-contained example

0.2.1

• Fixed linting issues.

0.2.0

• Restructured API, renamed liblsl to native_liblsl.

0.1.2

• Updated native_assets_cli to 0.12.0
• Updated native_toolchain_c to 0.9.0
• Removed spurious api.dart file (might come back later)

0.1.1

• Added missing meta dependency

0.1.0

• Restructured everything to be a bit more modular
• There's now an API for all the basic LSL functions 🥳🎈
• Started adding docs
• Test includes pushing and pulling a sample

0.0.2

• Android support 🎉

0.0.1

• Initial release
• Native compilation confirmed working on Windows, OSX and iOS. Will be testing Linux and Android soon.